feat: anthropic messages endpoint

api-reference/authentication.mdx

@@ -30,15 +30,25 @@ will also fail.
 <RequestExample>
 
   ```bash cURL with Bearer
-  curl 'https://api.edgee.ai/v1/projects' \
-  -H 'Authorization: Bearer <token>'
+  curl 'https://api.edgee.ai/v1/chat/completions' \
+  -H 'Authorization: Bearer <token>' \
+  -H 'Content-Type: application/json' \
+  -d '{
+    "model": "openai/gpt-4o",
+    "messages": [{"role": "user", "content": "Hello"}]
+  }'
   ```
-  
+
   ```bash cURL with Basic Auth
-  curl 'https://api.edgee.ai/v1/projects' \
-  -u '<token>:'
+  curl 'https://api.edgee.ai/v1/chat/completions' \
+  -u '<token>:' \
+  -H 'Content-Type: application/json' \
+  -d '{
+    "model": "openai/gpt-4o",
+    "messages": [{"role": "user", "content": "Hello"}]
+  }'
   # The colon prevents curl from asking for a password.
   ```
-    
+
 </RequestExample>
 

api-reference/errors.mdx

@@ -53,10 +53,11 @@ Below is a summary of the HTTP status codes that Edgee API uses.
 
 <ResponseField name="code" type="string">
   One of the following error codes:
-  
+
   - `bad_model_id`: The model ID format is invalid
   - `model_not_found`: The requested model does not exist or is not available
   - `provider_not_supported`: The requested provider is not supported for the specified model
+  - `streaming_not_supported`: Streaming is only supported when using Anthropic provider for Messages API
 </ResponseField>
 
 <ResponseExample>
@@ -86,6 +87,15 @@ Below is a summary of the HTTP status codes that Edgee API uses.
   }
 }
 ```
+
+```json Streaming Not Supported
+{
+  "error": {
+    "code": "streaming_not_supported",
+    "message": "Streaming for Messages API is only supported when the model uses the Anthropic provider"
+  }
+}
+```
 </ResponseExample>
 
 ### 401 Unauthorized

api-reference/index.mdx

@@ -26,10 +26,39 @@ When an error occurs, the Edgee API responds with a conventional HTTP response c
 
 ## Rate Limiting
 
-Please note that the Edgee has its own rate limit technology to prevent abuse and ensure service stability. 
-If you exceed these limits, your requests will be throttled and you will receive a `429 Too Many Requests` response. 
+Please note that the Edgee has its own rate limit technology to prevent abuse and ensure service stability.
+If you exceed these limits, your requests will be throttled and you will receive a `429 Too Many Requests` response.
 Additionally, usage limits may be enforced based on your API key configuration.
 
+## API Formats
+
+Edgee supports two API formats to accommodate different use cases and provider preferences:
+
+<CardGroup cols={2}>
+  <Card title="OpenAI Format" icon="robot" color="#8924A6">
+    Use `/v1/chat/completions` for maximum flexibility. Works with all providers (OpenAI, Anthropic, Google, Meta, etc.).
+
+    **Model format:** `provider/model` (e.g., `anthropic/claude-sonnet-4.5`)
+
+    **Recommended for:** New integrations, multi-provider applications, and maximum flexibility.
+  </Card>
+
+  <Card title="Anthropic Format" icon="message-square" color="#8924A6">
+    Use `/v1/messages` for native Anthropic Messages API format. Only works with Anthropic provider.
+
+    **Model format:** `model` (e.g., `claude-sonnet-4.5`)
+
+    **Recommended for:** Migrating from Anthropic's API, using Anthropic SDK, or requiring Anthropic-specific features.
+  </Card>
+</CardGroup>
+
+**Choose the format that best fits your needs:**
+
+- **For new projects**: Use the OpenAI format (`/v1/chat/completions`) for maximum flexibility and multi-provider support.
+- **For Anthropic migrations**: Use the Anthropic format (`/v1/messages`) for seamless compatibility with existing code using Anthropic's SDK.
+
+Both formats provide the same underlying AI capabilities and support Edgee's features like token compression, edge tools, and observability.
+
 ## Features
 
 <Columns cols={2}>